Simulated moving bed (SMB) chromatography uses a series of valves and columns to mimic continuous countercurrent flow between a solid adsorbent and a liquid feed, allowing continuous rather than batch separation. It dominates this patent set because it was the original solution to separating p-xylene from other C8 aromatic hydrocarbons at industrial scale, and the foundational apparatus patents for that application are among the most-cited records in the dataset. The mechanics later carried over into preparative and biologic purification, which is why SMB claims span both petrochemical and pharmaceutical IPC classes.

Filing activity peaked in 2018 at 30 records in this dataset and declined toward single digits by the 2022 midpoint, with recent-year momentum flat across every major assignee tracked. That pattern suggests the core apparatus claims were substantially staked out in an earlier window rather than the field being in active growth. New filing activity is more likely to succeed around narrower operating-parameter or application-specific claims than broad new apparatus architectures.

The dataset's ranking is concentrated among a small group of assignees spanning petrochemical processing, industrial gas and catalysis companies, and academic research foundations, several of which co-file with named individual inventors. All of the major assignees tracked show zero filings in the most recent year, meaning current activity is better read from the co-assignee and citation patterns than from year-over-year filing counts. A full assignee ranking with family counts is rendered in the accompanying table.

The clearest under-claimed branches sit adjacent to the dense SMB apparatus core: cycle-time and desorbent-flow optimization for feedstocks other than xylenes, resin capacity scaling specifically for peptide and protein purification, and solvent-consumption reduction in preparative-scale runs. These areas have some filing presence in the dataset, visible through the C07K and B01J IPC counts, but nowhere near the density of the core B01D separation-process claims, which suggests room for narrowly scoped claims that build on rather than duplicate the foundational patents.

Not necessarily. Citation counts inside a searched corpus tend to favour older records simply because they have had more time and more subsequent filings to accumulate citations against them, so the most-cited patents in this dataset are from the earlier petrochemical SMB era. A patent with fewer citations but a recent filing date and narrow claims on, for example, biologic purification parameters may be more commercially relevant to a current freedom-to-operate check than the historically dominant record.